|Home | About | Journals | Submit | Contact Us | Français|
Quality of life measures are increasingly used to supplement objective clinical or biological measures of disease to assess the quality of service, the need for health care, the effectiveness of interventions, and in cost utility analyses. Their use reflects a growing appreciation of the importance of how patients feel and how satisfied they are with treatment in addition to the traditional focus on disease outcomes. In this respect, quality of life measures capture patients' perspectives of their disease and treatment, their perceived need for health care, and their preferences for treatment and outcomes. They are hailed as being patient centred. But the challenge in measuring quality of life lies in its uniqueness to individuals. Many of the existing measures of quality of life fail to take account of this by imposing standardised models of quality of life and preselected domains; they are thus measures of general health status rather than quality of life.
Questions arise as to whether such measures are truly patient centred and to what extent they actually represent the quality of life of individual patients or groups of patients. Do they simply describe a patient's health in terms of what health professionals or society believe constitutes quality of life for people who are ill, something that may include factors that have little relevance to or importance for patients?
This paper explores the extent to which standardised quality of life measures accurately quantify an individual patient's quality of life. It debates whether newer, individualised approaches, which allow patients to define their quality of life in relation to their goals and expectations, are more appropriate.
Although there is no single agreed definition of health related quality of life, it is usually regarded as existing relative to individual or cultural expectations and goals (box). The first paper in this series proposed a model of quality of life that accounted for the interaction between expectations and experience.1 While it seems reasonable to assume that there are some aspects of life that are of universal relevance to quality of life, the specific weights that individuals attach to these will differ between and in different cultures. Other aspects may be important only to the individual. For example, the first paper in this series considered how the variations in expectations of health that exist between groups and individuals will have an impact on measuring quality of life. The interactions between all these aspects (generic and individual) will also vary between individuals.2 Moreover, these factors and their inter-relationships are unlikely to remain static over time.3 Values and priorities change in response to life circumstances, such as a life threatening illness, and experience, such as ageing or adapting to a chronic illness. Viewed in this way, both the determinants and evaluations of quality of life are highly specific to an individual.
Attempts to quantify and compare quality of life across different populations of patients using standardised generic measures have been confounded by the “disability paradox.” Patients who clearly have significant health and functional problems or intrusive symptoms do not necessarily have quality of life scores that seem commensurate with their health. In one study more than half of patients with moderate to severe disabilities reported having an excellent or good quality of life despite experiencing severe difficulties performing daily tasks, being socially isolated, and having limited incomes and benefits.8 Patients who have had transplants and patients having haemodialysis and peritoneal dialysis who reported a wide variety of health problems were more likely to rate themselves as “very happy” than the general population,9 and patients with neoplasms rated their quality of life in the top quarter of the World Health Organization's quality of life questionnaire (WHOQOL) across all life domains: this was better than all other groups of patients including those attending a family planning clinic.10
Quality of life is determined by
These discrepancies are replicated for individual patients as well. Several studies have shown that there is a disparity between patients', doctors', and relatives' ratings of the patient's quality of life11,12 or have suggested that doctors are unsuccessful in identifying aspects of disease and treatment that are important to patients.13,14 The implications of these findings for the use of proxies to measure patients' quality of life will be discussed in the next paper in the series.15
These data provide evidence that generic factors are individually weighted and that there are other factors important in quality of life that are not included in standardised measures. They also suggest that quality of life is a dynamic construct that alters in response to illness.
The lack of a widely agreed definition of health related quality of life means that many existing measures do not have any underlying theoretical conceptualisation of quality of life. Those few measures that are based on a theoretical model, such as the patient generated index,16 the repertory grid,17 and the World Health Organization's measures (the 100 item quality of life questionnaire (WHOQOL-100)6 and the 26 item questionnaire (WHOQOL-BREF)), are not widely used. This has led those who develop questionnaires to use a variety of sources for content. Many earlier questionnaires were based on health professionals' definitions of what was relevant. Some were based on reviews or adaptations of existing scales (such as the European quality of life measure (EuroQol),18 the McMaster health index,19 and the medical outcomes study 36 item short form health survey (SF-3620)). Few researchers directly asked patients about which factors they thought constituted quality of life. When they did involve patients, they asked about the impact of illness on people's lives or behaviour (for example, the Nottingham health profile21 and the sickness impact profile22) but not about the important things in people's lives. There is a danger that the most widely used measures do not address what is important to patients in determining their quality of life. For example, the most important factor influencing the quality of life of patients with cancer attending one outpatient clinic was reported to be that they were unable to find a parking place each time they went to the clinic; this was not addressed by any of the measures used.23 Further evidence for the limitations of measures in capturing what is important or relevant to patients comes from qualitative and survey research of the quality of life of different groups of patients.7,24 When the domains identified from these studies are compared with the domains captured by some of the most widely used measures (table), it is clear that although there is some overlap in the generic factors included, there are factors that are important to patients that are not captured by these measures. Additionally some factors may be redundant or irrelevant to patients.
Quality of life has many distinct but related determinants, some of which are captured by separate domains in measures. Scoring involves either recording the quality of life for each domain separately (profile measure) or combining the results from all domains to give a composite score (index measure). Meaningful interpretation of the results would be easier if there were some estimate of the relative importance of each of these domains. For example, pain is included in the physical domain of most measures but the importance attached to it varies across cultures.25
Quality of life measures have approached this problem in different ways. Some measures do not include any weighting assuming that patients find it impossible to put comparative values on important life domains, such as their family relationships and their ability to work. Other measures have used weights derived from the general population or from patients who have been asked to value the range of health states included in a questionnaire. Patients' responses to the questions are then valued according to these weights. However, these weights are unlikely to represent the values of individual patients. Different groups of patients attach a range of weights to the same domains (known as between patient variation),26 and the weights patients attach to the same domains at different periods in their treatment and recovery also change (known as within patient variation).27 In the same way that the determinants of quality of life are specific to individuals, the importance attached to those determinants will be influenced by an individual's expectations and aspirations as well as by their own belief system, their cultural belief system, and sociodemographic factors, such as age, sex, socioeconomic status, education, geographical location, and marital status. A true assessment of quality of life can only be achieved using weights for individual patients.
Using measures that are not patient centred can result in a number of problems. If they do not cover domains that are important to individual patients they may not be valid measures for those patients. Thus, standardised measures (in which the questions and range of answers are predetermined and the same for all patients) may measure something distinct from the quality of life of individual patients. In Bowling's study there were discrepancies between the free responses that patients made about the areas of their life that were most affected by disease and those elicited using “prompt cards”7; this suggests that results obtained using standardised measures may not capture a patient's quality of life.
If such measures do not capture the quality of life of individual patients they are unlikely to be responsive to change after treatment because they may not be measuring what is important to the patient and their scores may be difficult to interpret.
Measures that are not patient centred differ in content and the weights or importance they apply to different domains. Thus, significantly different scores may be obtained after the same intervention in the same patients. The SF-36 and the EuroQol measures, both of which are standardised measures, have produced contradictory results in the same patients when the effectiveness of cosmetic surgery has been assessed.28 This clearly has implications for determining the effectiveness of interventions, the relative quality of services, and the allocation of resources.
Research into quality of life measures should focus on
Quality of life can be measured in a patient centred way using individualised measures (box). Although less widely used than standardised measures, individualised measures are receiving increasing attention. However, they have their own problems. Firstly, some patients have difficulty understanding the system of direct weighting; this limits their use as self completed questionnaires29 among patients who are very sick or among those who have compromised concentration spans. Secondly, patients may not readily volunteer some factors that are important to them, particularly those related to mood,30 and the information that an individual is willing to volunteer may change over time.31 Finally, because of its individualised nature, the interpretation and analysis of some of the data are complex. This can make the comparison of groups of patients, or change within individuals over time, difficult, although many of the measures take account of this by using specific questions about change. Further work is needed to refine and evaluate these individualised measures.
Many of the most widely used measures of quality of life are limited in their ability to capture the quality of life of individual patients. These limitations result from the structure and content of the measures, the ways in which they were developed, and their systems of weighting. Some of these problems can be overcome by using individualised measures but these have their own problems which need further attention. A compromise may be to use recently developed standardised measures, which not only are sufficiently broad to include most facets of life important to any patient but which also use direct weighting systems; this should result in an individualised assessment of a patient's quality of life (box). The extent to which such measures reflect an individual's quality of life requires further assessment, and the clinical utility and interpretability of these measures also need to be established.
We thank our colleagues in the Interdisciplinary Research Group in Palliative and Person Centred Care at King's College London, in particular Peter Robinson, Barry Gibson, Stanley Gelbier, Robert Dunlop, Julia Addington-Hall, Lalit Kalra, Alan Turner-Smith, who have participated in discussions and commented on an earlier draft of this work.
This is the third in a series of five articles
Series editors: A J Carr, I J Higginson, P G Robinson
Funding: AJC's post is funded by the Arthritis Research Campaign.
Competing interests: None declared.
The articles in this series are from Quality of Life, which will be published by BMJ Books next year